Review Reports - Molecular Insights into Adsorption and Diffusion Mechanism of N-Hexane in MFI Zeolites with Different Si-to-Al Ratios and Counterions

Round 1

Reviewer 1 Report

See attached file

Comments for author File: Comments.pdf

Author Response

Dear editors and reviewers,

I am very grateful to your comments for the manuscript. We have carefully taken the reviewers′ comments into account and provided responses to each of the points raised by the reviewers. In addition, we also made a more detailed analysis of the data of the whole manuscript. All the altered passages have been highlighted. Please refer to the present manuscript text.

The comments

Part A (Reviewer 1)

The reviewer′s comments: Adsorption Isotherms

The authors report adsorption isotherms (Loading vs. hexane pressure) for systems at 298 and 425K with H⁺. Na⁺. and K⁺. For H⁺ at 298K we see three clusters of isotherms: one contains the three lowest SARs has a saturation limit ca. 40 and a half-saturation pressure value about 10(-9) kPa; the second, with SAR 30 and 50, has values near 32 and 10(-8); and the third with SAR 100 and 300 has values near 28 and 10(-6). Readings from the supplied graphs are highly approximate. Oddly, the untreated silicate has values in the second cluster. At the higher temperature the half-saturation values shift to higher pressure and the saturation limits are reduced.

Isotherms for Na (+) and K (+) are presented in Figure 4. The saturation limits are a bit smaller than for H (+), but the shift of half-saturation pressures upward at the higher temperature persists. The simple clustering of isotherms seen for H (+) is changed considerably

The authors′ answer:

Based on the reviewer's comments, we found that this part lacked information on data presentation, so we referenced the reviewer's text and revised accordingly (see lines 114-120 in manuscript). In addition, the value of untreated silicate-1 in the second group is explained and relevant literature is also referred ^[1] (see lines 122-124 and see Johannes A. Lercher et al work in science in manuscript). This literature shows that high-intensity proton environment can stabilize molecules, reduce reaction barrier and improve reaction activity, which can be used as the explanation for higher saturated limit of zeolites with low SARs (12.47, 5.86 and 2.43) than that of silicate-1 in this work.

[1] Pfriem N, Hintermeier P H, Eckstein S, et al. Role of the ionic environment in enhancing the activity of reacting molecules in zeolite pores[J]. Science, 372.

The reviewer′s comments: Mean-square displacement of n-hexane

The zeolite structures have straight (STR) and sinusoidal (SIN) channels of diffusion. There may be different mobilities in STR and SIN channels. Figure 3 displays averaged mean-square displacement (MSD) over time of hexane in the several substrates. At 298 the lowest SAR materials display greater mobility for hexane. Mobility increases dramatically for these systems as the temperature is raised to 423K. Relative diffusion constants, derived from the slope of the MS displacement plotted against time, are larger for these low-SAR species.

The diffusion behavior for Na (+) and K (+) is more complex. The discussion of Figure 7 (lines 204215) is in my opinion not sufficiently detailed. Mention is made of changes in distribution of hexane between STR and SIN channels, but a more quantitative evaluation of that distribution would be much more informative than the complex diagrams. (See comment below on Figure 2.)

The authors′ answer:

We have completed the supplement to the discussion in Figure 7 (see lines 246-260 in manuscript).

Additional work was done to perform quantitative evaluation of n-hexane distribution in STR and SIN channels. Firstly, we reworked the original data of density distribution (Part of the data in Figure 1) and tried to estimate densities of different channels quantitatively. Subsequently, we tried to analyze the distribution of n-hexane in STR and SIN channels using the contour surface, which is much clearer and more informative than the previous figures (see. Figures 2). However, in the analysis of NaMFI and KMFI systems, we found that this contour surface figure is only applicable to the case of the system with strong nonlocality of density distribution, such as HMFI. For highly localized systems, such as NaMFI and KMFI, the selection of contour surface in some straight channels is subjective, and we cannot judge which straight channel contour surface is appropriate. Finally, we choose to highlight the density distribution of n-hexane in the straight channels with the blue shadows. The modified diagrams are shown in Figures 3-8 (see also Figures 2, 5 and 6 and Figures S1-S3 in manuscript). In addition, we found that the previous analysis was not detailed and clear enough, and the overall modification was made for the density distribution of n-hexane (see lines 134-140, 202-208 and 222-224 in manuscript).

Figure 1 Partial original data of density distribution of n-hexane molecules.

Figure 2 Density distribution of n-hexane molecules in the one straight channel contour map.

Figure 3. Density distribution of n-hexane on silicate-1(a) and HZSM-5 with different SAR of 300 (b), 100 (c), 50 (d), 30 (e), 12.47 (f), 5.86 (g) and 2.43 (h) at 298 K. The X-axis direction (vertical direction in picture) represents the sinusoidal channels; the Y-axis direction (horizontal direction in picture) represents the straight channels (blue shadow area). green region represents the lowest-energy, preferred adsorption positions for n-hexane molecules.

Figure 4. Density distribution of n-hexane on silicate-1 (a) and NaZSM-5 with different SAR of 300 (b), 100 (c), 50 (d), 30 (e), 12.47 (f), 5.86 (g) and 2.43 (h) at 298 K. The X-axis direction (vertical direction in picture) represents the sinusoidal channels; the Y-axis direction (horizontal direction in picture) represents the straight channels (blue shadow area). The green region represents the lowest-energy, preferred adsorption positions for n-hexane molecules.

Figure 6. Density distribution of n-hexane on silicate-1 (a) and HZSM-5 with different SAR of 300 (b), 100 (c), 50 (d), 30 (e), 12.47 (f), 5.86 (g) and 2.43 (h) at 423 K. The X-axis direction (vertical direction in picture) represents the sinusoidal channels; the Y-axis direction (horizontal direction in picture) represents the straight channels (blue shadow area). The green region represents the lowest-energy, preferred adsorption positions for n-hexane molecules.

Figure 6. Density distribution of n-hexane on silicate-1 (a) and KZSM-5 with different SAR of 300 (b), 100 (c), 50 (d), 30 (e), 12.47 (f), 5.86 (g) and 2.43 (h) at 298 K. The X-axis direction (vertical direction in picture) represents the sinusoidal channels; the Y-axis direction (horizontal direction in picture) represents the straight channels (blue shadow area). The green region represents the lowest-energy, preferred adsorption positions for n-hexane molecules.

Figure 7. Density distribution of n-hexane on silicate-1 (a) and NaZSM-5 with different SAR of 300 (b), 100 (c), 50 (d), 30 (e), 12.47 (f), 5.86 (g) and 2.43 (h) at 423 K. The X-axis direction (vertical direction in picture) represents the sinusoidal channels; the Y-axis direction (horizontal direction in picture) represents the straight channels (blue shadow area). The green region represents the lowest-energy, preferred adsorption positions for n-hexane molecules.

Figure 8. Density distribution of n-hexane on silicate-1 (a) and KZSM-5 with different SAR of 300 (b), 100 (c), 50 (d), 30 (e), 12.47 (f), 5.86 (g) and 2.43 (h) at 423 K. The X-axis direction (vertical direction in picture) represents the sinusoidal channels; the Y-axis direction (horizontal direction in picture) represents the straight channels (blue shadow area). The green region represents the lowest-energy, preferred adsorption positions for n-hexane molecules.

The reviewer′s comments: Notes on anisotropy of diffusion

The anisotropy in diffusion is important, and I am glad to see that the authors report the components (Figures S4 and S5) at least for Na(+). But it can hardly be correct to say ”high temperatures weaken the interaction between Na(+) and hexane.” (Lines 221-222) nor can it be said without further context that “…high temperature limits the diffusion of n-hexane molecules in the straight channel.” Conceivably high temperatures tip a balance so that transport in SIN channels is enhanced over that in STR channels. Supplementary material referring to the anisotropy of diffusion of hexane in Na ZSM-5 (Figures S4 for T=298 and S5 for T=423) show MSD(ZZ), MSD(XX), and MSD(YY), x being aligned with SIN channels and y aligned with STR channels. MSD(ZZ) is always small - why is that? At 298 MSD is largest (ca 90 A2) after 700 psec for SAR = 100, and dominated by D(YY). Raising the temperature to 425K reduces the MSD to about 55 units, but D(XX) and D(YY) make roughly equal contributions to that smaller value. Raising the temperature also drastically reduces the MSD for the system with SAR = 300. I do not understand how this can be.

Figure 7 contains some other mysteries. Increasing temperature increases the diffusion in Silica in a perfectly normal way. For the Al-bearing systems the diffusion behavior shifts in puzzling ways. I do not think the discussion addresses all the oddities. For example in high temperature K(+) with SAR=300 shows structure – could it be a second slow process only accessible at high T?

The authors′ answer:

The incorrect description of the original manuscript (lines 221-222) has been corrected (see lines 266-268 in manuscript).

There are no channels in the Z-axis direction, which is why the diffusion coefficient in this direction (ZZ) is basically zero (see lines 262-264 in manuscript).

The unusual phenomenon (1. the remarkable decrease of the MSD for the Na(+) obtained for high temperature case with SAR = 300 and 2. the existence of a second slow process of the MSD for the K(+) system at high temperature with SAR=300) were explained (see lines 258-260 and 276-282 in manuscript).

The reviewer′s comments: Isosteric heat of adsorption

Figure 8 shows heats of adsorption vs. loading (molecules of n-hexane/cell) for the range of SAR at two temperatures. Panels a and b refer to HZSM-5, and for Silicate-1 show a roughly linear increase in strength as loading increases. It appears that the binding is weaker at the higher temperature. (This could be a consequence of differences of heat capacity correction terms (See also lines 266-267)). Fitting the data to regression lines would allow a more confident statement. Binding curves for the range of SAR values seem noisier; regression would also diagnose the departure from linearity.

Data for Na(+) is still more irregular, and not complete. The graphs for K(+) show further structure/ scatter. Is there a way to explain why data is so erratic? The mention (lines 267-269) that there may be distinct distribution can possibly be tested. (See comment below on Figure 2.)

The authors′ answer:

The R2 values of linear fitting data of adsorption heats are shown in table 1 (see also Table S1 in supporting). The relevant description is provided in the revised manuscript (see lines 301-310 in manuscript).

For Na+ systems, the lower SAR, the lower adsorption amount, which may cause the data to look incomplete, but in fact the data is complete. We have made a more detailed analysis of adsorption heat data for Na+ and K+ the system (see lines 328-330 and 341-351 in manuscript) and attempted to explain why the data of K+ systems is so erratic (see lines 341-351 in manuscript) and are trying to use the results of quantum chemical calculation to reveal the essence of complex host-guest interaction mechanism in the following research.

Table S1. The R² values of linear fitting data of adsorption heats

Zeolite/Temperature	298K	423K
Silicate-1	0.94004	0.9788
HMFI(300)	0.65442	0.84637
HMFI(100)	0.90097	0.82822
HMFI(50)	0.92652	0.93038
HMFI(30)	0.96873	0.98369

The reviewer′s comments: Interaction between n-hexane and the M(+)Alumino-silicate zeolites

The authors display data for interaction between n-hexane and the zeolite framework for a range of SAR and two temperatures at a range of hexane pressures, for HZSM-5 (Figure 9), NaZSM-5 (Figure 10), and KZSM-5 (Figure 11). For example, the color-coded lines shown in panel c (SAR=50, T=298K) of Figure 9 referring to P(E) - the probability of specific values of interaction energy –show two clusters. The low-pressure cluster (< 5x10(-7) kPa) refers to relatively strong interaction around about -24.5 kcal/mol and the higher-pressure (> 10(-5) kPa) cluster refers to weaker interaction, at most -24 kcal/mol. There is considerable overlap of the several distributions. At the higher temperature (panel g) the clustering is less evident and there seems to be a shift toward lower adsorption energy. There is at least a hint that the two peak values may refer to the two channels STR and SIN. The Na (+) system graphs are broadly similar though the clustering has vanished, while the K (+) systems shows quite different behavior in detail. Still the shift of the P(E) peak to weaker hexane-zeolite interaction at the higher temperature persists. If one knew the occupancy of SIN and STR channels, perhaps these details could be rationalized.

The authors′ answer:

According to the reviewer's description, we performed a more detailed analysis of the interaction energy curves of the three systems HMFI, NaMFI and KMFI (see lines 316-322 and 348-351 in manuscript).

The reviewer′s comments: Distribution of sites with H(+)

The authors present radial distribution functions (RDFs) for the separation between H(+) and specific protonation sites, for the most acidic system with SAR ca 2.5; the RDF for sites independently established as favored. (But are the sentences “It can be seen… sorption and reaction [8,20, 22].” strictly consistent? What of T9?)

The authors′ answer: For T9 site, it is a very promising n-hexane adsorption site although it may not be easily replaced by aluminum (see lines 382-383 in manuscript).

Further comments on figures:

Line 124: (Figure 1) Rather than (per&cell) one might write “hexane molecules per cell.” A common symbol (triangles of varied color) is used; different shapes as in Figure 8 would ease reading of the graphs. This applies also to Fig 4.

Line 127: (Figure 2) Sinusoidal and straight diffusion channels are shown, with green circles depicting adsorption sites. Remarks are made about preference for sinusoidal or straight channels (lines 112 – 122), but to this reader it is not easy to infer the relative occupation of channels. Is a numerical measure of population in channels (NSTR/NSIN) not available in the MD results? This comment applies to all the related figures 5, 6, S1, S2, S3, S8, and S9.

Line 256: (Figure 8) Plotting the (negative) heat of adsorption usually places the most negative value (in this case -26.0 kcal/mol) at the bottom of the vertical axis.

The authors′ answer:

Figures in former manuscript (Line 124 and Line 256) have been revised as recommended by the reviewer, as displayed in Figures 9-11 (see also Figures 1, 4 and 8 in manuscript). The reviewer's comments to Figures in former manuscript (e.g. Line 127) have been explained and modified above (see 2. The reviewer′s comments).

Figure 9. Adsorption isotherm of n-hexane molecules on silicate-1 and HZSM-5 with different SAR (300, 100, 50, 30, 12.47, 5.86 and 2.43) at 298 K (a) and 423 K (b).

Figure 10. Adsorption isotherm of n-hexane molecules on NaMFI (a, b) and KMFI (c, d) with different SAR (∞, 300, 100, 50, 30, 12.47, 5.86 and 2.43) at 298 K (a, c) and 423 K (b, d).

Figure 8. Isosteric heat of n-hexane molecules on HZSM-5 (a, b), NaZSM-5 (c, d) and KZSM-5 (e, f) zeolites with typical SAR of 300, 100, 50 and 30 at 298K (a, c, e) and 423K (b, d, f).

Discussion

The first sentence (lines 313 makes mention of the distribution of n-hexane in STR and SIN channels. It would seem important to find a simple way to characterize that distribution (relative population within channels). This has already been encouraged above, and is I believe the best way to improve the paper.

The authors′ answer: The reviewer's comments have been explained and modified above (see 2. The reviewer′s comments).

Minor issues of usage

Line 157: replace “extend” with “extent”

(The sentence containing this word in the revised version has been deleted)

Line 213: replace “siltcate-1” by “silicate-1”

(Has been modified, line 255)

Line 314: replace “promoted due to” by “enhanced by”

(Has been modified, line 392)

Line 374: omit “suffic

Dear editors and reviewers,

The comments

Part A (Reviewer 1)

The reviewer′s comments: Adsorption Isotherms

The authors′ answer:

[1] Pfriem N, Hintermeier P H, Eckstein S, et al. Role of the ionic environment in enhancing the activity of reacting molecules in zeolite pores[J]. Science, 372.

The reviewer′s comments: Mean-square displacement of n-hexane

The authors′ answer:

We have completed the supplement to the discussion in Figure 7 (see lines 246-260 in manuscript).

Figure 1 Partial original data of density distribution of n-hexane molecules.

Figure 2 Density distribution of n-hexane molecules in the one straight channel contour map.

The reviewer′s comments: Notes on anisotropy of diffusion

The authors′ answer:

The incorrect description of the original manuscript (lines 221-222) has been corrected (see lines 266-268 in manuscript).

There are no channels in the Z-axis direction, which is why the diffusion coefficient in this direction (ZZ) is basically zero (see lines 262-264 in manuscript).

The reviewer′s comments: Isosteric heat of adsorption

The authors′ answer:

Table S1. The R² values of linear fitting data of adsorption heats

Zeolite/Temperature	298K	423K
Silicate-1	0.94004	0.9788
HMFI(300)	0.65442	0.84637
HMFI(100)	0.90097	0.82822
HMFI(50)	0.92652	0.93038
HMFI(30)	0.96873	0.98369

The reviewer′s comments: Interaction between n-hexane and the M(+)Alumino-silicate zeolites

The authors′ answer:

The reviewer′s comments: Distribution of sites with H(+)

The authors′ answer: For T9 site, it is a very promising n-hexane adsorption site although it may not be easily replaced by aluminum (see lines 382-383 in manuscript).

Further comments on figures:

Line 256: (Figure 8) Plotting the (negative) heat of adsorption usually places the most negative value (in this case -26.0 kcal/mol) at the bottom of the vertical axis.

The authors′ answer:

Figure 9. Adsorption isotherm of n-hexane molecules on silicate-1 and HZSM-5 with different SAR (300, 100, 50, 30, 12.47, 5.86 and 2.43) at 298 K (a) and 423 K (b).

Figure 10. Adsorption isotherm of n-hexane molecules on NaMFI (a, b) and KMFI (c, d) with different SAR (∞, 300, 100, 50, 30, 12.47, 5.86 and 2.43) at 298 K (a, c) and 423 K (b, d).

Figure 8. Isosteric heat of n-hexane molecules on HZSM-5 (a, b), NaZSM-5 (c, d) and KZSM-5 (e, f) zeolites with typical SAR of 300, 100, 50 and 30 at 298K (a, c, e) and 423K (b, d, f).

Discussion

The authors′ answer: The reviewer's comments have been explained and modified above (see 2. The reviewer′s comments).

Minor issues of usage

Line 157: replace “extend” with “extent”

(The sentence containing this word in the revised version has been deleted)

Line 213: replace “siltcate-1” by “silicate-1”

(Has been modified, line 255)

Line 314: replace “promoted due to” by “enhanced by”

(Has been modified, line 392)

Line 374: omit “sufficiently” (redundant)

(Has been modified, line 465)

Line 385: omit passage (385 – 389) bearing on databases

(Has been modified, line 476-480)

iently” (redundant)

(Has been modified, line 465)

Line 385: omit passage (385 – 389) bearing on databases

(Has been modified, line 476-480)

Author Response File: Author Response.docx

Reviewer 2 Report

After reading the manuscript, I regard it a good piece of work, worthy of publication in this journal. Before doing so, I have some questions to make to the authors. The first one is related to the absence of any experimental result that bear out the reported isotherms. I think some experimental results should be included, from the authors or from literature references to support the theoretical isotherms. At least, some indication is to be given that renders the reader trust in these results. The second question is related to the mention in different parts of the manuscript that the incorporation of the alkali cations stops n-hexane diffusion. I think this must be better explained because my impression is that alkali cations in MFI slows diffusion but it does not stop it. Additionally, it seems strange to me that isosteric heat of alkali MFI are not very different from HMFI, despite the remarkable differences reported in diffusivity. I think these results of allosteric heat should be checked by the authors.

Author Response

Dear editors and reviewers,

The comments

Part B (Reviewer 2)

The reviewer′s comments:

The first one is related to the absence of any experimental result that bear out the reported isotherms. I think some experimental results should be included, from the authors or from literature references to support the theoretical isotherms. At least, some indication is to be given that renders the reader trust in these results.

The authors′ answer:

We have referenced some experimental and simulation literatures to support the theoretical isotherms ^[1-3] (see also line 104-105 in manuscript). The type of adsorption isotherm of n-hexane in this manuscript is consistent with that in the experimental n-hexane isotherm ^{[2, 3]}. The accuracy of simulation also can be demonstrated by CBMC simulations ^{[1, 4]}.

[1] H, Jobic, N, et al. Influence of Isotherm Inflection on the Loading Dependence of the Diffusivities of n-Hexane and n-Heptane in MFI Zeolite. Quasi-Elastic Neutron Scattering Experiments Supplemented by Molecular Simulations [J]. The Journal of Physical Chemistry B, 2006, 110(5):2195–2201.

[2] Ruthven D M , Kaul B K . Adsorption of n-hexane and intermediate molecular weight aromatic hydrocarbons on LaY zeolite [J]. Industrial & Engineering Chemistry Research, 1996, 35(6):2060-2064.

[3] Jeroen, Lannoeye, Tim, et al. Adsorption and separation of n-hexane and cyclohexane on the UiO-66 metal-organic framework [J]. Microporous & Mesoporous Materials the Offical Journal of the International Zeolite Association, 2014.

[4] Vlugt T, Krishna R, Smit B. Molecular simulations of adsorption isotherms of linear and branched alkanes and their mixtures in silicalite [J]. The Journal of Physical Chemistry B, 1999, 103(7).

The reviewer′s comments:

The second question is related to the mention in different parts of the manuscript that the incorporation of the alkali cations stops n-hexane diffusion. I think this must be better explained because my impression is that alkali cations in MFI slows diffusion but it does not stop it.

The authors′ answer:

Based on the reviewer's description, we have modified the original words (e.g. restricts n-hexane diffusion) describing diffusion process into more accurate words (e.g. slows hexane diffusion) (see also lines 275 in manuscript).

The reviewer′s comments:

Additionally, it seems strange to me that isosteric heat of alkali MFI are not very different from HMFI, despite the remarkable differences reported in diffusivity. I think these results of isosteric heat should be checked by the authors.

The authors′ answer:

The results of equivalent adsorption heat have been carefully checked and are correct. The isosteric heat are closely related to the structure of the molecule and the interaction mode between the molecule and different active sites. The values take into account not just the acid site but the local environment around the adsorption site. Therefore, HMFI, not alkali MFI, show a roughly linear increase in strength as loading increases, although there is little difference in isosteric heat value between HMFI and alkali MFI system. In fact, we also get that benzene has similar adsorption energy on protons (31 kcal·mol^-1) and Na+ cations (32 kcal·mol^-1) in DFT calculation.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Followup comments on Molecular Insights into adsorption and diffusion

The authors have responded conscientiously to my comments. Small usage errors have been corrected or eliminated, and the figures 1, 4, and 8 are easy to read owing to their use of distinctive point symbols. Many other improvements have been incorporated, including explanations of puzzling features of the diffusion.

The authors still maintain that ”high temperature weakens the interaction between Na(+) and n-hexane molecules.” (Lines 310-311) This cannot be; there is no temperature dependence in the potential energy. At higher temperature one might expect kinetic energy to overcome the attraction between hexane and the substrate, easing diffusion. See also lines 381-382.

The authors have worked to clarify the channel density diagrams, Figures 2, 5, 6, S1, S2, S3, S8, and S9. Interpretation is still difficult for me. For example, the authors state “These results show that the HZSM-5 zeolites with SAR less than 50 [favor] straight channels.” I cannot see this. And I do not see a reply to my question from the first round “Is a numerical measure of population in channels (NSTR/NSIN) not available in the MD results?” What I have in mind is a direct count of the number of molecules in Straight channels and the number of molecules in Sinusoidal channels, which may change in time as diffusion occurs, and may depend on temperature. Is this count very difficult?

Minor issues:

Lines 247-256:

in the passage “…slows diffusion the diffusion of n-hexane…” delete the word underlined.

Later, “It should note…” can be “it should be noted…”

The sentence beginning “This phenomenon…” should be rephrased for clarity.

Line 278:

R2 values need not be reported to more than two significant figures.

Line 279:

The sentence beginning “The distribution of n-hexane…” needs rephrasing.

Author Response

Dear editors and reviewers,

We greatly appreciate your further comments on the manuscript. We have carefully taken the reviewer′ comments into account and provided responses to each of the points raised by the reviewer. All the altered passages have been highlighted. Please refer to the present manuscript text and attach file.

Author Response File: Author Response.pdf

Round 3

Reviewer 1 Report

Remarks on Molecular insights into adsorption and diffusion (2^nd revision)

I am happy to see the report of population of hexanes in channels, which I think clarifies the distribution very simply.

My thanks to the authors for patience and cooperation.

Minor points:

Line 53: Replace “sitting” by “sited”

Line 97: Does “scatters” really mean “points”?

Line 207: Delete “And”

Line 312: In the sentence beginning “Figures 8c and 8d…” please replace “refer” by referring” and replace “slightly” by “slight”

Author Response

Dear editors and reviewers,

We greatly appreciate your further comments on the manuscript. Minor points have revised. Please refer to the present manuscript text.

Minor points

Line 53: Replace “sitting” by “sited”

(Has been revised, line 53 in manuscript)

Line 97: Does “scatters” really mean “points”?

(This expression has been seen in the literature, but the “scatters” has been changed into the “points” , line 97 in manuscript)

Line 207: Delete “And”

(Has been deleted, line 207 in manuscript)

Line 312: In the sentence beginning “Figures 8c and 8d…” please replace “refer” by referring” and replace “slightly” by “slight”

(Has been revised, line 312 in manuscript)

Author Response File: Author Response.pdf